The formation and nature of the generative cell wall and the detachment mode of the generative cell from the intine in Polystachia pubescens were observed by LM and TEM. Vesicles evenly positioned within the phragmoplast fuse to form a cell plate that divides the microspore into the generative and vegetative cell. This cell plate consists of callose. Before the generative cell leaves the intine, however, the callose is completely resorbed and is not replaced by any other substance. The generative cell becomes detached from the intine by moving towards the centre of the pollen grain. A constriction formed thereby gives the generative cell a bulb-like appearance and leads ultimately to the generative cell being pinched off. Plasma-filled vesicles originating from the generative cell remain between the infine and the plasma membrane of the vegetative cell.

No abstract

Clear spruce wood specimens were subjected to in-situ tensile failure tests parallel to the grain in order to study fracture initiation and propagation. Fracturing of the specimens was carried out inside a SEM with a bending and a tension device, respectively. The bending and the tension specimens consisted of two to three growth rings. In all specimens, fracture propagated in a stepwise manner and the earlywood zones fractured prior to the latewood zones. The fracture initials comprised parallel-to-grain cracks at the earlywood-latewood borders and perpendicular-to-grain cracks in wood rays at a growth ring border. The parallel-to-grain cracks were probably due to transverse tensile strains and were frequently arrested or became discontinuous at wood rays. Wood rays were shown to be structural points of weakness as well as reinforcements depending on die direction of load. The role of the bordered pits remained ambiguous.

Studying fracture initiation and propagation in wood at a microscopical level should provide much information about fracture mechanics and cell wall structure. To this end, a method was developed to carry out fracturing of wood in a scanning electron microscope. A bending and a tension module were built to fit the chamber of a SEM. Both devices are designed to observe tension failure. Load was implemented parallel to the grain of wood samples, Norway Spruce (Picea abies L. Karst.). Crack initiation and propagation were recognized and recorded in most samples. Although Scanning Electron Microscopy proved to be very useful, it also poses some difficulties. The electron beam can strongly influence the fracture behaviour and morphology of wood. However, electron beam damage-induced fractures in wood have very characteristic features and, therefore, can be distinguished from natural fractures. If certain precautionary measures are observed, electron beam damage can be avoided. IntroductionWood strength and fracture behaviour vary within and between species. These differences cannot be completely explained by a macroscopical description, but need to be studied on a microscopical level (Bariska 1994;Kucera and Bariska 1982). Scanning electron microscopy (SEM) would, thereby, be the method of choice since it allows the examination of relatively large specimens and does not require elaborate specimen preparation. The examination of fracture surfaces in the SEM has, so far, demonstrated how anatomical features influence the fracture morphology of wood and has also provided much information about the ultrastructure of the cell wall (Kucera and Bariska 1982;Sell and Zimmermann 1993;Zimmermann 1993). However, little is known about crack development on the microscopical and ultrastructural level because fracturing has mainly been done outside the SEM. Studying the actual fracture process in the SEM should, therefore, provide even more information about the fracture mechanism in wood. In-situ fracturing of wood in the SEM, however, poses some difficulties. The main problems are chamber size, specimen preparation, vacuum quality and, above all, electron beam damage (Hoffmeyer and Hanna 1989). The present paper demonstrates a method that allows to study the crack initiation and crack propagation in wood during tension and bending failure tests in the SEM. It is also concerned with the identification and avoidance of electron beam damage. The results of the failure tests will be presented in a follow-up paper.

Clear spruce wood specimens with 30, 20, 10, and 5° slope of grain were subjected to in-situ tensile failure tests in order to study fracture initiation and propagation. Fracturing of the specimens was carried out inside a SEM with a bending and a tension device, respectively. The bending and the tension specimens consisted of two to three growth rings. Fracture process in 30°-, 20°-, and 10°-specimens was quite uniform but differed from that of 5°-specimens. There was no difference between bending and tension specimens. With one exception, the specimens fractured in a stepwise manner. Cracks initiated at wood rays in the latewood and were arrested in the adjoining earlywood. This is attributed to the greater stiffness of the latewood tracheids and the capacity of the thin-walled earlywood tracheids to deform thereby arresting cracks. In 30°-, 20°-, and 10°-specimens, the main fracture mode was cleavage along the grain due to tensile stress perpendicular to the grain. Latewood tracheids showed intrawall fracture while earlywood tracheids showed longitudinal transwall failure. In the 5°-specimens, tensile stress parallel to the grain prevailed.